GAMME, a meta-model to unify data needs in simulation modeling (WIP)

摘要

Complex system studies, especially in the aerospace context, where systems are too expensive and too critical to allow real experimentation, make extensive use of simulation. A drawback of the simulation approach is that, in general, aerospace systems studied are so large that their simulations are quite difficult to develop. Indeed, such a simulation will generally involve simulation components coming from various sources and sharing information. There, one may see two opposite constraints. On the one hand, the simulation components are developed independently from each other (or potentially preexist) and hence need to manage on their own the data that they need. On the other hand, overall data consistency has to be ensured in the simulation, especially for shared data. An interesting way to solve this issue consists in structuring the data, i.e. defining a datamodel that will be used as a support to specify all operations relevant to data (access, storage, ...): one may see this approach as the application of model-driven methods to the central point of data in a simulation. This paper presents GAMME, a home-made data-centered framework for the development of large-scale simulations, as well as uses of this framework in an ONERA simulation example.